Probability-based prediction query algorithm

Advances in wireless communication and ubiquitous mobile networks have resulted in location-based services (LBSs) contributing significantly to entertainment services for mobile users by utilizing their location data. However, for the request of a mobile user with a limited time frame, such as “where is the gas station nearest to me? The anonymous server should return the optimal result to me within a few minutes,” the anonymous server immediately generates the cloaking region based on the user's current location information and the k-anonymity requirement privacy level to ensure location privacy protection when receiving the LBS request in most existing approaches. Unfortunately, mobility by the mobile user could cause the smaller cloaking region generated by the server. The existing algorithm would return the sub-optimal cloaking region. This problem prompted us to develop a novel probability-based prediction query (PBPQ) algorithm. On the server side, our method firstly predicts the probability of the mobile user being on each road, followed the final probability of the user entering the cloaking region. Our proposed algorithm can generate the optimal cloaking region generated within the limited time required by users. The experimental results show that our PBPQ algorithm can predict the optimal cloaking region effectively at the highly successful rate of more than 90%. Thus, the PBPQ algorithm can decrease the size of the cloaking region considerably while maintaining good QoS.